use tau

Author: Lazy Programmer

rl/comparing_explore_exploit_methods.py

@@ -27,8 +27,7 @@ def sample(self):
     return np.random.randn() / np.sqrt(self.lambda_) + self.predicted_mean
 
   def update(self, x):
-    # assume tau is 1
-    self.lambda_ += 1
+    self.lambda_ += self.tau
     self.sum_x += x
     self.predicted_mean = self.tau*self.sum_x / self.lambda_